Embed Size (px)
Citation preview
HOW TO ACCOMMODATE INCREASED
BANDWIDTH REQUESTS OF ADVANCED
APPLICATIONS
Josef Vojtěch et al. Optical networks department, CESNET
10th CEF Networks workshop
September 3rd 2019, Praha
Lada Altmannová, Sarbojeet Bhowmick, Ondřej Havliš,
Michal Hažlinský, Tomáš Horváth, Jaroslav Jedlinský,
Václav Kubernát, Petr Munster, Jan Kundrát, Jan Radil,
Pavel Škoda, Martin Šlapák, Radek Velc, Rudolf Vohnout
Optical networks department, CESNET
Co-authors
CEF Networks Workshop
2004 1st CEF - 2019 10th CEF
Established by Stanislav Šíma
However repeating of CEF is allowed by
your interest and participation
3
2004 1ST CEF - 2019 10TH CEF
- developments
4
Long haul serial speed: 10 Gbps -> 600/800 Gbps (31/34% annually ~ 1.7-1.8 /2 years)
Moore’s: twice in two years, twice in 18 months (60% annually)
Nielsen's law of internet bandwidth: A high-end user's connection speed grows
by 50% per year.
From fibre acquisition to submarine spectrum
From lighting to open lighting, open spectrum
From IP to advanced applications
CESNET2
5
2004
Building elements: dark fibres, optical transmission equipment, routers, switches
No DWDM
3000 km DF lines (included 350 km single fibre)
Only single CBF (SANET)
New network function: provide E2E circuits
Increasing number of customer premises fibre
connected
CESNET2
6
2019
Transmission systems: Commercial + Open Line System Czech light™, 100 Gbps speeds
5830 km (including 1820 km, single fibre cost effective and
also required by advanced application)
CBF triangle since 2006 (ACONet, CESNET, SANET)
Advanced photonic services based on optical fibre
CESNET2 – Precise Time and Frequency
7
Fibres shared with data
Deployed over 2000 km of fibre
Time transfer + distribution
TTAs, WR, both w/ DWDM optics
Comparison UTC(TP), UTC(BEV)
Research Institute of Geodesy
Topography and Cartography, Faculty
of electrical engineering
Czech Metrology Institute
ELI
CESNET2 - Time and Frequency
8
Coherent optical frequency transmission
500 -> over 1000 km of lines, further 200km to be
commissioned
Sources based in Praha, Brno, Wien. Optical
clock under development in Brno Olomouc.
Dedicated all-optical sub band 800/400 GHz (incl.
1540.56 and 1542.1 nm)
This Presentation
This presentation illustrates the near shortage of spectrum in C band
Presents spectrum use by advanced network services: delivery of ultra-stable optical frequency / time or performing distributed sensing sharing telecom fiber
Focus on alternatives to 1530-1565 nm C band – S, L
Case study of 119 km line in C, L and S bands
9
10
Spectrum Usage
Obsolete – 96 ch. per 10 Gbit/s – 0.96 Tbpss (OOK)
Mature – 96 ch. per 100 Gbit/s – 9.6 Tbpss (DP-QPSK)
Deployed – 96 ch. per 200 Gbit/s – 19.2 Tbps (DP-16QAM)
𝑂𝑆𝑁𝑅𝐴𝑆𝐸= Pout - NF - G - 10 log(N) + 58
source: cisco.com
11
Capacity sharing trial 400 Gbps
ACOnet (AT), CESNET (CZ), SANET (SK),
Crossborder fibres 519 km a 134 dB, Open Line System Czech Light
(100 GHz grid)
56 Gbaud system 200/300/400 Gbps, required 59 GHz of optical spectrum
(only 48 channels, 19.2 Tbps)
High attenuation avg. 26.8 dB / span
Very high OSNR required
300 Gbps achieved
Spectrum Usage
12
fluctuation ~130 ns (temp. changes 12 deg C)
TDEV 8.7 ps / 500 s, @ 1552.52 nm
Smotlacha, V., Kuna, A., Mache, W., "Time Transfer in Optical Network," Proceedings of the 42nd Annual Precise Time and Time Interval Systems and Applications Meeting, Reston, Virginia, November 2010
Precise Optical Time Transfer
Univ. Paris 13, LNE-SYRTE, RENATER, 150 km
Lopez et al., Opt. Express 18, 16849 (2010)
@1542.1 nm
Ultra-stable Optical Frequency Transfer
13
Long-term measurement of the stability and shape deviation
of the containment buildings
Two PWR reactors ea 1 GWe, protected by its containment building
Precise measuring methods based on Fiber Bragg Gratings strain gauges
Vojtěch J. et al „Joint stable optical frequency and precise time transfer over 406 km of shared fiber lines – Study“, In proc 40th TSP, 2017
@1540.56 nm
14
Ultra-stable Optical Frequency Transfer
Side product of ultra-stable optical frequency dissemination
25 Apr 2016 – ML 4.1, epicentre located 20 km SW from Vienna ■ @1540.56 nm
M. Cizek et al., "Transfer of stable optical frequency for sensory networks via 306 km optical fiber link,"
2016 European Frequency and Time Forum (EFTF), 2016
Optical Fibre Seismology
15
Submarine optical link: Malta Sea earthquake - Sept 2017
G Marra, C Clivati, R
Luckett, A Tampellini, J
Kronjäger, L Wright, A
Mura, F Levi, S Robinson,
A Xuereb, B Baptie, D
Calonico „Ultrastable
laser interferometry for
earthquake detection with
terrestrial and submarine
cables“, Science, 361,
486-49 (2018)
70% of the Earth’s ocean bottom. Submarine lines are far quiet (up to 40 dB) compared to terrestrial ones
Seismometers installation difficult and expensive - over 1 million km of submarine cable already installed
Potential of important application: By detecting underwater earthquakes close to their epicenter,
life-saving time could be gained in a tsunami warning ■ @1542.1 nm
Optical Fibre Seismology
16
Fiber Optic Gyroscopes
C. Clivati et al „A Large Area
Fiber Optic Gyroscope on
multiplexed fiber network“,
Optics letters. 38. 1092-4.
10.1364/OL.38.001092.
20 km2, sensitivity about (10−8 rad/s)
Fibre shared with data transmissions
@1542.1 nm
Fiber Optic Gyroscope
17
Backscatter
is measured
while pulse
propagates
Based on signal intensity change
localisation of event is possible
in order of tens of meters
Transmission fibre
Vibrations
Phase sensitive-OTDR
MÜNSTER, P. et al „Phi- OTDR signal amplification“. In Proc. SPIE
9506, Optical Sensors 2015. SPIE, 2015. p. 1-9.
@1550.1 nm
18
Distributed Fibre Acoustic Sensing
Bypassing telco transmission system – within S band
NIKHEF – VSL 2014, 137 km
Koelemeij J. et al “Methods for data, time and ultrastable
frequency transfer through long-haul fiber-optic links”
@1470+1490 nm
Precise Optical Time Transfer
19
T/F transfer + distribution
Dedicated all-optical sub
band 800/400 GHz (incl.
1540.5 and 1542.1 nm)
20
Time and Frequency Infrastructure
S band C band U band
1450-1530 1530-1565 1570-1605 1605-1675
L band
21
Optical Clock Interconnection
S band C band U band
1450-1530 1530-1565 1570-1605 1605-1675
L band
Optical clock based on trapped
and cooled single ion 40Ca+
under development
Direct output at: 729 nm
-> 1458 nm
Distance: 119km, 29 dB
C 1570 nm L
22
Bidirectional EDF Amplifiers Comparison
Need reciprocal/bidirectional path to cancel
slow changes 𝜏AB = 𝜏BA
Bidirectional amplification??
Hi gain medium + feedback
We are trying to avoid it!!
G2R1R2 < 1
R composes from Rayleigh backscattering
and reflections from splices, connectors etc.
Only with limited gain up to 20-22 dB
Even over lossy spans:
24, 27.7, 26 and 28.6 dB
23
Challenge – Bidirectional Amplification
Monitor and avoid unwanted oscillations, keep max possible gain
Bidi No-Lase EDFA
Raman - quite ineffective (500mW to 8 dB of gain)
Semiconductor Optical Amplifier
Will require some gain stabilization e.g.
Holding beam injection
Estimate max available gain 16 - 18dB
24
S Band Amplification
Vojtech, J., Radil, J. and Smotlacha, V., (2015). Semiconductor Optical Ampl.ifier with Holding Beam Injection for Single
Path Accurate Time Transmission. JTh2A.78. 10.1364/CLEO_AT.2015.JTh2A.78.
Propposed design I
25
Optical Clock Interconnection
L band penalty max 0.02 dB/km totally 2.4 dB
S band according G.652D max 0.31 dB/km at
1460 nm, fiber attenuation about 36.9 dB
System Attenuation
[dB]
DWDM data systems 25
C band bidi 32.2
L band bidi 34.6
S band bidi 39.9
4 km 7 km
26
Optical Clock Interconnection
L band penalty max 0.02 dB/km totally 2.4 dB
S band according G.652D max 0.31 dB/km at
1460 nm, fiber attenuation about 36.9 dB
System Attenuation
[dB]
DWDM data systems 25
C band bidi 35.2
L band bidi 37.6
S band bidi 41.4
Propposed design II
C band spectrum is becoming rare
L band will provide some additional portion
Than spatial multiplexing
The SCL line
Low attenuation and minimal reflections
Careful selection of passives components, all ideally fused together
S band SOAs should be fine for CW
Backup - Thulium doped fluoride fibre based amplifiers
Passives partially deployed in the field
27
Summary
Jan Gruntorád, Helmut Sverenyak
Martin Míchal, Jakub Mer, Josef Verich, Václav Novák
This work was supported partially by the Ministry of Education, Youth and Sport of
the Czech Republic as part of the project "E infrastructure CESNET - modernization",
reg. nr. CZ.02.1.01/0.0/0.0/16_013/0001797
Acknowledgement
Time and frequency = quantities we are able to measure with the highest accuracy
Represent ideal way how to measure tiny effects
(Radio)astronomy: VLBI, SKA
Precise tests of fundamental physics:
Constancy of fundamental constants
Detection of gravitational wave
Tests of special & general relativity
Credits: Newbury14, Barr10
Why Precise Time and Frequency?
30
Credit: Schnatz14
Credits: Colorado, Timetech, Droste13
CV GNSS (GPS, GALILEO, GLONASS, …) accuracy 3 – 50 ns
GNSS PPP (Precise Point Positioning) 0.1 ns
TWSTF 0.1 ns
Satellite RF Based Transfer
31
Clocks Are Improving
‘microwave’ atomic clocks ‘optical’ atomic clocks
32
Using telecom transmission system - lambdas
White Rabbit time-transfer experiment between Espoo and Kajaani in Finland
2013 Collaboration: MIKES/CSC/FUNET ~ 900 km
https://www.ohwr.org/projects/white-rabbit/wiki/mikes
33
Using telecom transmission system – lambdas
Ebenhag S. et al “Coherent Optical Two-Way Frequency Transfer in a Commercial
DWDM Network” PTTI 2016
